Consistency of augmentation graph and network approximability in contrastive learning

Open in new window